0

I am reading a technical paper and I don't quite understand some statistical terms such as:

  1. ..."Although the asymptotic mean of $z(\phi)$ is zero, it is possible for the score statistic derived by this theory to have means of order $O(n_{ij}^{-1/2})$"... what does $O(n_{ij}^{-1/2})$ means here?
  2. ...$E[z(\phi)]=O(n_{0}^{-3/2},n_{1}^{-3/2})$, that is the bias to order $O(n_{0}^{-1/2},n_{1}^{-1/2})$ is zero. What does the $O(n_{0}^{-1/2},n_{1}^{-1/2})$ mean?

Where $z(\phi)$ is the z-score from standard normal distribution in terms of $\phi$ and $n$ is sample sizes in integers.

David Z
  • 1,618
  • 1
    Does this help? https://en.wikipedia.org/wiki/Big_O_notation – Stephan Kolassa Sep 28 '22 at 12:31
  • 3
    related: https://stats.stackexchange.com/questions/207264/root-n-consistent-estimator-but-root-n-doesnt-converge/207281#207281 https://stats.stackexchange.com/questions/306521/o-p-notation-and-consistent-estimators/306577#306577 – Christoph Hanck Sep 28 '22 at 12:47

0 Answers0